Dysregulation of ossification-related miRNAs in circulating osteogenic progenitor cells obtained from patients with aortic stenosis

• •

  The osteogenic processes are involved in mechanisms in CAVD. Ossification-related miRNAs may play an important role in osteogenic differentiation from bone marrow stromal cells and the osteogenic processes in CAVD.

• •

  The present study has suggested that dysregulation of ossification-related miRNAs such as miR-106a, miR-148a, miR-204, miR-211, miR-30a, miR-31 and miR-424 may be related to differentiation into COPCs and thereby play an important role in the pathogenesis of CAVD.

• •

  Levels of ossification-related miRNAs and COPCs count were changed by surgical treatments and may be novel post-operative biomarkers.

CAVD (calcific aortic valve disease) is the most common valvular heart disease in the aging population and is associated with significant morbidity and mortality [1]. CAVD represents a progressive pathological process progressing from slight thickening of the aortic valve without obstruction of blood flow to aortic valve sclerosis and finally to AS (aortic stenosis) with severe calcification of the valvular leaflets, reduction in valve motion and obstruction of blood flow [1]. Previous studies have shown CAVD to be mediated by a chronic inflammatory disease process, with many similarities to the pathological process underlying atherosclerosis which includes inflammatory cell infiltrates, lipoproteins, extracellular bone matrix proteins and bone minerals [2]. A recent report implicates osteogenic processes as key mechanisms in CAVD [3]. In particular, circulating osteogenic progenitor cells are involved in heterotopic ossification in the valve tissues [4,5]. Gössl et al. [6] have reported that high numbers of COPCs (circulating osteogenic progenitor cells) indicate accelerated progression of CAVD in humans. However, little is known of the mechanisms of the cellular and molecular components and processes of CAVD.

miRNAs are the most abundant family of small non-coding RNAs and these regulate mRNA translation of target genes through the RNAi pathway [7,8]. We have reported previously that expression of inflammation- or longevity-related miRNAs in endothelial progenitor cells is involved in the pathogenesis of coronary atherosclerosis [9,10]. Some studies have suggested that miRNAs play an important role in osteogenic differentiation from bone marrow stromal cells and mesenchymal cells [11,12]. In particular, it has been reported that ossification-related miRNAs, including miR-106a, miR-148a, miR-204, miR-211, miR-30a, miR-31 and miR-424 act as important endogenous negative or positive regulators of osteogenesis of human mesenchymal progenitor cells and bone marrow stromal cells [11,13,14]. However, it remains unknown whether expression of ossification-related miRNAs is related to osteogenic differentiation of circulating progenitor cells in patients with CAVD.

The aim of the present study was to determine whether ossification-related miRNAs are expressed with osteocalcin in COPCs obtained from patients with CAVD compared with controls. In addition, the present study evaluated the effects of surgical treatments on expression of ossification-related miRNAs and osteocalcin levels.

The present study was approved by the ethics committee of the Iwate Medical University School of Medicine (H26-46), and written informed consent was obtained from all subjects. A total of 46 patients with severe clinical and echocardiographic AS were referred for pre-operative clinical examination. The severity of AS was determined by standard echocardiographic methods. Severe AS was defined as symptomatic AS or an aortic valve area of ≤1.0 cm² or an aortic valve area index of ≤0.6 cm²/m² or mean aortic gradient >60 mmHg or peak aortic velocity >5 m/s [15]. Patients were excluded from the study if they had clinical signs of acute coronary syndrome, acute infection, severe renal failure (serum creatine levels over 3 mg/dl) or rheumatoid disease, or if they were suspected of having a malignant or primary wasting disorder.

As controls, 46 subjects were identified from a list of outpatients attending our university hospital for treatment of hypertension, dyslipidaemia or arrhythmia, and showing no evidence of AS on history or physical examination. Subjects were selected from the case loads of clinicians who were not related to this study in order to reduce bias and were stratified according to sex and age to match each AS case.

Fasting peripheral blood was collected from patients with AS in the morning after an overnight fast for baseline data before surgical treatment and again 2 and 12 weeks after surgical treatment [AVR (aortic valve replacement) or TAVI (transcatheter aortic valve implantation)]. Fasting peripheral blood was also collected from controls in the morning after an overnight fast.

PBMCs (peripheral blood mononuclear cells) were isolated from heparinized blood samples obtained from all subjects by Ficoll–Paque density gradient centrifugation. PBMCs were incubated with FITC-conjugated anti-CD133 mAb (monoclonal antibody) (Miltenyi Biotec), PerCP (peridinin–chlorophyll protein complex)–Cy5.5 (N,N′-dicarboxypentyl-indodicarbocyanine-5,5′-disulfonic acid)-conjugated anti-human CD34 mAb (Becton Dickinson), PE (phycoerythrin)-conjugated anti-human KDR (kinase insert domain-containing receptor) mAb (R&D Systems), APC (allophycocyanin)-conjugated anti-human osteocalcin mAb (R&D Systems).

Cell enumeration by FACS (Becton Dickinson) was performed immediately after the staining with all samples protected from light. To eliminate clumping, doublet discrimination was utilized as a part of the FACS procedure. From each sample, a limit of 10⁷ events was established. Events were gated for monocytes, live cells, dead cells and subcellular debris using forward-scattered and side-scattered lights. Live cells were gated for FITC-conjugated CD133-positive and CD34–PerCP–Cy5.5-negative and KDR–PE-positive cells. This subset was gated further for osteocalcin-positive cells for enumeration of COPCs. The numbers of COPCs among 10⁶ cells were counted using a FACScan analyser. Isotype-matched irrelevant control IgG (PerCP, PE or APC) was used as a control (Becton Dickinson).

Osteocalcin levels in COPCs were measured using a FACScan flow cytometer (Becton Dickinson) and shown as MFI (mean fluorescence intensity).

Total RNA including the small RNA fraction was extracted from isolated PBMCs using a mirVana™ Paris miRNA isolation kit (Ambion), according to the manufacturer's instructions.

Extracted RNA from isolated COPCs was initially reverse-transcribed using a High Capacity cDNA Archive Kit (Applied Biosystems) and then amplified in a 10 μl PCR volume and primer set for amplification of human miR-106a (Assay ID 000578), miR-148a (Assay ID 000470), miR-204 (Assay ID 000508), miR-211 (Assay ID 000514), miR-30c (Assay ID 000419), miR-31 (Assay ID 001100), miR-424 (Assay ID 000604) and U6 (Assay ID 001093) using TaqMan® miRNA assays (Applied Biosystems) according to the manufacturer's recommended protocol. The amplification steps consisted of denaturation at 95°C, followed by 40 cycles of denaturation at 95°C for 15 s and then annealing at 60°C for 1 min. All reactions were carried out on the 7500 real-time PCR system (Applied Biosystems) using the TaqMan® Universal PCR master mix and Assays on demand (Applied Biosystems). Relative quantification was carried out using the ΔΔC_T (threshold cycle) method for recurrent compared with primary with U6 as an endogenous control and fold changes were calculated for each gene [16]. Replicates with C_T > 40 were excluded. The assay was run in triplicate for each case to allow for assessment of technical variability. To account for PCR amplification of contaminating genomic DNA, a control without reverse transcriptase was included. To improve the accuracy of real-time RT (reverse transcription)–PCR for quantification, amplifications were performed in triplicate for each RNA sample.

Cultured PBMCs from five patients with AS (random sampling) were used for an in vitro model transfected with miR-106a, miR-148a, miR-204, miR-211, miR-30c, miR-199a, miR-31 or miR-424 mimic and inhibitor. The protocol for cultured PBMCs was as described previously [9]. For these miRNA mimic and precursor transfections, PBMCs were transfected with pre-miR™ miR-106a, miR-148a, miR-204, miR-211, miR-30c, miR-199a, miR-31 or miR-424 precursor molecules (Ambion, 20 ng/ml) and anti-miR™ miR-106a, miR-148a, miR-204, miR-211, miR-30c, miR-199a, miR-31 or miR-424 inhibitors (Ambion, 20 ng/ml) using the NeoFx transfection agent (Ambion). Those cells were used for experiments 24 h after transfection and were exposed at each passage to 0.1 μM t-BHP (t-butyl hydroperoxide) (Sigma–Aldrich) or PBS for an additional 48 h. Levels of these miRNAs and osteocalcin MFI were measured using the above methods.

AVC (aortic valve calcification) was quantified from MDCT (multidetector computed tomography) scans according to the method originally described by Clavel et al. [17]. In brief, the degree of AVC was semi-quantitatively classified into no calcification and calcification according to the AVC values of 2065 AU (arbitrary units) in men and 1274 AU in women as the cut-off values to define AVC, and was evaluated in aortic valve short axis images at both systolic and diastolic phases [17]. The percentage of AVC was expressed as a ratio of 100× an area of AVC/a total area of AV. The representative cases of MDCT in patients with severe AVC and moderate AVC are shown in Figure 1. The MDCT images were randomly evaluated by two independent expert observers (K.T. and T.N.) who were blinded to all clinical characteristics.

All values are presented as means±S.D. Kolmogorov–Smirnov analysis was performed to assess data distribution. An unpaired Student's t test was performed for normally distributed data, and non-parametric Mann–Whitney test was performed where this was not appropriate. Statistical analysis of categorical variables was also carried out using χ² analysis and Fisher exact analysis. After 2 or 12 weeks of surgical treatments, comparisons between the two groups (AVR or TAVI) were analysed by two-way repeated-measures ANOVA for normally distributed variables. When applicable, significant differences were analysed further with Dunnett's post-hoc tests. Pearson's correlation coefficients were used to examine the relationship between the number of COPCs, miRNAs, osteocalcin protein levels and the degree of AVC. A value of P<0.05 was considered statistically significant.

Baseline characteristics of the study populations are shown in Table 1. There were no significant differences in age, percentage of males and BMI (body mass index) between the AS group and controls. There were significant differences in other parameters between the two groups (P<0.05) (Table 1). There were significant differences in some baseline and clinical characteristics between the AVR and TAVI groups (Table 1). The data from echocardiography and pressure studies are shown in Table 2. There were no findings of aortic calcification in the controls. There were significant differences in haemodynamic parameters between the AVR and TAVI groups (Table 2).

The number of COPCs (CD133-positive/CD34-negative/KDR-positive/osteocalcin-positive cells) was higher in the AS group than in the controls (55.0±14.7 compared with 16.8±10.8 per 10⁶ events; P<0.01). There was no significant difference in the number of baseline COPCs between the AVR and the TAVI groups (53.6±12.8 compared with 56.2±16.5 per 10⁶ events; not significant). In addition, there was no significant difference in the number of baseline COPCs between AS patients with CAD (coronary artery disease) (n=11) and those without CAD (n=35) (57.1±16.8 compared with 55.0±14.6 per 10⁶ events; not significant).

There was no difference in average C_T of U6 between the AS group and the controls (26.8±2.3 compared with 27.9±1.9; not significant). Levels of miR-30c were higher in the AS group than in controls (4.05±1.62 compared with 0.96±0.76; P<0.01) (Figure 2A). Levels of miR-106a, miR-148a, miR-204, miR-211, miR-31 and miR-424 were lower in the AS group than in the controls (miR-106a, 0.92±0.53 compared with 5.17±2.05; P<0.01; miR-148a, 0.94±0.61 compared with 4.93±2.44; P<0.01; miR-204, 0.95±0.53 compared with 4.07±1.68; P<0.01; miR-211, 0.92±0.62 compared with 4.43±1.91; P<0.01; miR-31, 0.84±0.50 compared with 4.07±1.97; P<0.01; miR-424, 0.97±0.66 compared with 5.20±2.31; P<0.01 (Figure 2A). Osteocalcin MFI was higher in the AS group than in the controls (17.59±3.52 compared with 14.53±2.89; P<0.01) (Figure 2B). There were no significant differences in levels of ossification-related miRNAs and osteocalcin MFI between AS patients with CAD and those without CAD (miR-106a, 0.68±0.44 compared with 0.94±0.53; miR-148a, 0.82±0.46 compared with 0.99±0.67; miR-204, 0.86±0.56 compared with 1.00±0.56; miR-211, 1.13±0.68 compared with 0.91±0.60; miR-31, 0.61±0.25 compared with 0.90±0.56; miR-424, 0.89±0.48 compared with 1.08±0.63; osteocalcin MFI, 18.35±3.97 compared with 17.67±3.42 not significant).

Levels of miR-30a were positively correlated with the number of COPCs and osteocalcin MFI. On the other hand, levels of miR-106a, miR-148a, miR-204, miR-211, miR-31 and miR-424 were negatively correlated with the number of COPCs and osteocalcin MFI (Table 3). The number of COPCs and levels of miR-106a, miR-148a, miR-204, miR-211, miR-30c, miR-31 and miR-424 were statistically correlated with haemodynamic profiles including aortic valve area, mean aortic gradient and peak flow velocity (Table 4). The degree of AVC was weakly positively correlated with miR-30c levels and negatively correlated with miR-106a and miR-424 levels (miR-30c compared with AVC, r=0.48, P<0.01; miR-106a compared with AVC, r=−0.43, P<0.05; miR-148a compared with AVC, r=−0.29, P=0.11; miR-204 compared with AVC, r=−0.10, P=0.60; miR-211 compared with AVC, r=−0.21, P=0.24; miR-31 compared with AVC, r=−0.23, P=0.21; miR-424 compared with AVC, r=−0.43, P<0.05).

The number of COPCs and osteocalcin MFI were decreased 2 and 12 weeks after AVR or TAVI (Figures 3A and 3B). The AVR group showed decreased miR-30c levels 2 and 12 weeks after AVR and increased levels of miR-106a, miR-148a, miR-204, miR-211, miR-199a, miR-31 and miR-424 3 and 12 weeks after AVR (Figure 4). The TAVI group showed decreased miR-30c levels 2 weeks after AVR and increased levels of the remaining miRNAs 2 weeks after AVR, whereas levels of these miRNAs did not change 12 weeks after TAVI (Figure 4). Fold changes in the number of COPCs, osteocalcin MFI and ossification-related miRNA levels were higher in the AVR group than in the TAVI group (Figure 4).

Cultured PBMCs were transfected with ossification-related miRNA mimic and inhibitor followed by challenge with t-BHP after 48 h. Levels of miR-106a, miR-148a, miR-204, miR-211, miR-30c, miR-31 and miR-424 expression were higher in the mimic-transfected cells than in the mock-transfected cells (Figure 5A). On the other hand, levels of miR-106a, miR-148a, miR-204, miR-211, miR-30c, miR-31 and miR-424 expression were lower in the inhibitor-transfected PBMCs than in the mock-transfected PBMCs (Figure 5A). Cultured PBMCs transfected with miR-106a, miR-148a, miR-204, miR-211, miR-199a, miR-31 and miR-424 mimic showed a reduction in levels of osteocalcin MFI compared with mock-transfected cells (Figure 5B). PBMCs transfected with miR-106a, miR-148a, miR-204, miR-211, miR-199a, miR-31 and miR-424 inhibitor showed an increase in levels of osteocalcin MFI compared with mock-transfected cells (Figure 5B). On the other hand, PBMCs transfected with miR-30c mimic showed an increase in levels of osteocalcin MFI compared with mock-transfected cells (Figure 5B). PBMCs transfected with miR-30c inhibitor showed a decrease in levels of osteocalcin MFI compared with mock-transfected cells (Figure 5B).

The main findings of the present study are: (i) the number of COPCs was higher in the AS group than in controls; (ii) levels of ossification-related miRNAs were shown to differ between the AS group and controls and to be statistically correlated with the number of COPCs; (iii) surgical treatment with TAVI or AVR generated markedly changed levels of ossification-related miRNAs and number of COPCs; and (iv) in vitro study showed that ossification-related miRNAs resulted in negative or positive regulation of osteocalcin maturation.

The cellular origin of ectopic AVC has been unclear for a long time, despite this being considered to be a major cause of CAVD. Some reports have shown that COPCs were released from bone marrow and then recaptured into the bone microenvironment [3,18]. It has also been reported that COPCs which are osteocalcin-, bone alkaline phosphate- or type I collagen-positive PBMCs participate in ectopic osteogenic injury responses during fracture and valve ossification [4,19]. We have shown an increase in the number of COPCs exhibiting positive CD133, KDR and osteocalcin and negative CD34 in patients with AS compared with controls. In addition, the number of COPCs is positively correlated with the degree of AVC. It has been reported that an increase in COPCs is associated with unstable CAD and tends to predict further events [20]. However, there were no significant differences in the number of COPCs between AS patients with CAD and those without CAD. It has been speculated that a surge of COPCs may be released from bone marrow in response to aortic calcification. In agreement with our results, it has been reported that a high number of COPCs was associated with the severity of AS [6]. It has also been reported that COPCs (type I collagen-positive/CD45-positive) appear in human aortic valves during several phases of AS including the fibroproliferative and neovascularization phases [5]. These observations suggest that an increase in the number of COPCs may be involved in the progression of AVC.

Accumulating evidence suggests that some ossification-related miRNAs, such as miR-106a, miR-148a, miR-204, miR-211, miR-30c, miR-31 and miR-424 are involved in the process of heterotopic ossification modulating osteogenesis via up-regulation or down-regulation of these ossification-related miRNAs [11,13,14]. The present study demonstrates up-regulation of miR-30c levels and down-regulation of miR-106a, miR-148a, miR-204, miR-211, miR-31 and miR-424 levels in patients with AS compared with controls. Overexpression of miR-30c induces heterotopic ossification in mesenchymal stem cells via expression of several osteogenic mediators, such as ITGB1 (integrin β1), CXCL12 (CXC chemokine ligand 12), FLT1 (Fms-like tyrosine kinase-1), CAMTA1 (calmodulin-binding transcription activator 1) or COL9A3 (collagen type IX alpha 3), and therefore acts as a positive regulator of osteogenesis [12,13]. On the other hand, it has been reported that miR-106a, miR-148a, miR-204, miR-211, miR-199a, miR-31 and miR-424 act as negative mediators of osteogenesis via repression of several osteogenic mediators, such as RUNX2 (Runt-related transcription factor 2), BMPR2 (bone morphogenetic protein receptor type 2), CBFB (core-binding factor beta), BMP2 (bone morphogenetic protein 2), BMP3 and BMP8B [12,13]. The present study has also shown that osteocalcin levels were higher in the AS group than in controls, and osteocalcin levels and AVC were correlated with levels of ossification-related miRNAs. Osteoblasts produce osteocalcin and incorporate it into the bone matrix [21]. Osteocalcin is released into the circulation from the matrix during bone resorption and is considered a marker of bone turnover [21]. Our in vitro study showed gain-of-function and loss-of-function approaches to ossification-related miRNAs, and transfection of each ossification-related miRNA into cultured PBMCs resulted in regulation of osteocalcin expression, suggesting that these ossification-related miRNAs may play an important role in osteocalcin maturation. These observations suggest that abnormal profiling of ossification-related miRNAs may play an important role in osteogenesis in patients with AS. In addition, the present study shows that the degree of AVC is positively correlated with miR-30c levels and negatively correlated with miR-106a and miR-424 levels in patients with AS. It may be a novel finding that dysregulated profiling of ossification-related miRNAs is related to the degree of AVC. It has therefore been suggested that dysregulated profiling of ossification-related miRNAs may be involved in the progression of aortic calcification in patients with AS. Although the present study has not evaluated the effect of medications such as statins on aortic calcification in patients with AS, a recent study has demonstrated that oxLDL (oxidized low-density lipoprotein) increased the valve calcification-related signalling pathway by increasing extracellular bone matrix protein, and atorvastatin also prevented any oxLDL-induced effects through the same pathway [22]. Further study will be needed to clarify the mechanism by which other miRNAs dysregulate osteogenesis in AVC among patients with AS.

Another important finding of the present study was that miR-106a, miR-148a, miR-204, miR-211, miR-199a, miR-31 and miR-424 levels increased after surgical treatment with TAVI or AVR. On the other hand, miR-30c levels and the number of COPCs were reduced after surgical treatment. There have been no reports concerning the effect of surgical treatment on ossification-related miRNA expression and COPC count. From these observations, it may be speculated that surgical treatment ameliorated CAV by activating valvular endothelial cells and inflammatory cells and releasing the aortic pressure gradient of AS, and then terminating the osteogenic processes via dysregulation of ossification-related miRNAs. The present study has also shown that changes in ossification-related miRNA levels, the number of COPCs and osteocalcin MFI after surgical treatment were greater in the AVR group than in the TAVI group. This result may be explained by procedural differences between AVR and TAVI. Although AVR eliminated CAVD completely, TAVI left residual CAVD tissue around the artificial cardiac valve after the procedure, which may continue to activate the osteogenic process via dysregulation of ossification-related miRNAs. It is therefore suggested that TAVI may not terminate the process of osteogenesis completely. These observations suggest that the remaining COPCs and miR-30c after TAVI may be related to the progression of biological valve sclerosis and result in a worsening of long-term prognosis. It is also speculated that an increase in COPCs via dysregulation of ossification-related miRNAs may play an important role as a poor prognostic factor in severe AS patients without AVR or TAVI.

Bioprosthetic valve degeneration is a major complication after the implantation of biological prosthetic valves [23]. It has been reported that the incidence of structural valve degeneration was 3.6% at 6 years and 5% at 10 years after implantation [24]. A study involving microscopic analysis of explanted bioprosthetic heart valves revealed the presence of moderate to severe calcification as well as CAVD [25]. A pathological study has demonstrated that bioprosthetic tissue valve calcification was similar to that of native valves [26]. These reports suggest that the osteogenic processes via dysregulation of ossification-related miRNAs may be involved in the pathogenesis of bioprosthetic valve degeneration. Further long-term follow-up studies will therefore be required.

The study was designed and conducted by Kan Takahashi, Yuji Takahashi, Mamoru Satoh and Yoshihiro Morino. Data analysis was performed by Kan Takahashi, Takuya Osaki, Takahito Nasu, Makiko Tamada, Motoyuki Nakamura and Yoshihiro Morino. The paper was written by Kan Takahashi, Mamoru Satoh and Yoshihiro Morino.

This study was supported by a Grant-in-Aid for General Scientific Research from the Japanese Ministry of Education, Science, Sports and Culture [grant number 25461138].

APC

allophycocyanin

AS

aortic stenosis

AU

arbitrary units

AVC

aortic valve calcification

AVR

aortic valve replacement

BMP

bone morphogenetic protein

CAD

coronary artery disease

CAVD

calcific aortic valve disease

COPC

circulating osteogenic progenitor cell

Cy5.5

N,N′-dicarboxypentyl-indodicarbocyanine-5,5′-disulfonic acid

KDR

kinase insert domain-containing receptor

mAb

monoclonal antibody

MDCT

multidetector computed tomography

MFI

mean fluorescence intensity

oxLDL

oxidized low-density lipoprotein

PBMC

peripheral blood mononuclear cell

PE

phycoerythrin

PerCP

peridinin–chlorophyll protein complex

TAVI

transcatheter aortic valve implantation

t-BHP

t-butyl hydroperoxide